Image Alignment

Image Alignment restores the camera position and orientation at the time of capture, as well as the sparse 3D scene structure, based on original images, geolocation information (GNSS and optional GCP), and optional initial camera interior orientation parameters. The image alignment result can be used in the subsequent DOM generation workflow.

Note: When running this function, an NVIDIA discrete graphics card no lower than the GTX 1660 series is recommended.

Parameters

• Optimize Interior Orientation (all selected by default): Supported interior orientation parameters include camera focal length (f), principal point (cx, cy), radial distortion parameters (k1, k2, k3), and tangential or decentering distortion parameters (p1, p2). When no initial interior orientation is available, the internal self-calibration algorithm can estimate good interior orientation values. Calibrated camera interior orientation can also be used. If externally imported interior orientation parameters are highly accurate, clear this option so that the parameters remain unchanged during adjustment.

• Estimate Boresight Error (all cleared by default): Supported boresight parameters include GPS boresight error, namely offsets along the X, Y, and Z axes from GPS to camera, and IMU boresight error, namely heading, pitch, and roll angle offsets from IMU to camera. These boresight parameters are 0 by default, meaning no GPS or IMU boresight error. GPS boresight error can be estimated accurately only when GPS accuracy is high and accurate ground control points (GCPs) are available. Do not select this option when GPS accuracy cannot be guaranteed or no control points are available.